摘要
In this retrospective cohort of 18 066 Medicare fee-for-service beneficiaries ≥ 65 years with acute VTE (2015-2019), we compared apixaban, rivaroxaban, and warfarin using propensity score (PS) overlap weighting and intention-to-treat analyses across total, non-frail, and frail strata defined by a claims-based frailty index. After PS weighting (all SMDs < 0.10), apixaban versus warfarin was associated with a lower 1-year composite of recurrent VTE or death in the total population (189.1 versus 216.1 per 1000 PY; HR 0.86, 95% CI 0.77-0.96), with benefit evident in non-frail patients (114.5 versus 143.6; HR 0.78, 0.66-0.92) and attenuated in frail patients (330.9 versus 349.9; HR 0.93, 0.82-1.07). Recurrent VTE rates favored apixaban but were not statistically significant overall (40.5 versus 54.8; HR 0.86, 0.68-1.08). Major bleeding was lower with apixaban (19.4 versus 26.0; HR 0.73, 0.53-1.00). Home-time loss was reduced with apixaban compared with warfarin in the total population (mean 54.9 versus 67.0 days; RR 0.89, 0.83-0.96) and was directionally similar across both frailty strata. Rivaroxaban showed no clear advantage over warfarin for the composite (216.9 versus 216.1; HR 1.00, 0.90-1.11) or bleeding (HR 0.99, 0.73-1.32) and yielded greater home-time loss than apixaban (RR 1.08, 1.01-1.17). Overall, apixaban demonstrated the most favorable balance of effectiveness, safety, and patient-centered benefit, with patterns generally consistent across frailty levels.